返回Google圖書搜尋

Heteroskedasticity in Repeat Sales House Price Equations

出版	SSRN, 1998
URL	http://books.google.com.hk/books?id=Bp_mzgEACAAJ&hl=&source=gbs_api

註釋Several authors have attributed heteroskedasticity in repeat sales house price equations exclusively to length of time between sales. Goodman and Thibodeau (1995) developed a theoretical model relating heteroskedasticity in hedonic house price equations to dwelling age. Their model has the following implications for repeat sales equations? Properties must be older to be held longer, so heteroskedasticity observed in repeat sales equations could be caused primarily by dwelling age rather than by the time between sales. Alternatively, owner-occupied homes tend to be improved (by sellers, buyers, or both) at the time of sale. The shorter the time between sales, the less extensive are the improvements. Less extensive (undocumented) improvements translate into more accurate predictions of house prices and of subsequent appreciation rates. Consequently, the heteroskedasticity observed in repeat sales equations could also be related to the time between sales holding dwelling age constant. Empirical results using nearly 2,000 Dallas repeat sales from 1979 through 1993 indicate that both dwelling age and time between sales contribute to heteroskedasticity in repeat sales equations. The Davidian and Carroll iterative estimation technique is then used to incorporate both dwelling age and time between sales to estimate efficient repeat sales house price indexes.